Microbiome data integration workflow for population cohort studies

Package demo

# Get starting time
starting_time <- Sys.time()

################################################################################
# List of packages that we need
packages <- c("ggplot2", "mia", "miaViz", "dplyr", "tidyr", "scater",
              "ComplexHeatmap", "knitr")

# Get packages that are already installed installed
packages_already_installed <- packages[ packages %in% installed.packages() ]

# Get packages that need to be installed
packages_need_to_install <- setdiff( packages, packages_already_installed )

# Loads BiocManager into the session. Install it if it not already installed.
if( !require("BiocManager") ){
    install.packages("BiocManager")
    library("BiocManager")
}

# If there are packages that need to be installed, installs them with BiocManager
# Updates old packages.
if( length(packages_need_to_install) > 0 ) {
   install(packages_need_to_install, ask = FALSE)
}

# Load all packages into session. Stop if there are packages that were not
# successfully loaded
if( any(!sapply(packages, require, character.only = TRUE)) ){
    stop("Error in loading packages into the session.")
}

################################################################################
# Additional setup

# Set black and white theme for figures, and Arial font
theme <- theme_bw() + theme(text = element_text(family = "Arial"), 
                            panel.border = element_blank(), 
                            panel.grid.major = element_blank(),
                            panel.grid.minor = element_blank(), 
                            axis.line = element_line(colour = "black"))
theme_set(theme)

EuroBioC2023

• European Bioconductor Conference 2023
• Ghent, Belgium
• September 21, 2023; 13:30 (CEST)
• See the poster also (miaverse – microbiome analytics framework in SummarizedExperiment family)!

Presenter information

All authors are affiliated to Turku Data Science Group in University of Turku, Finland.

• Tuomas Borman – doctoral researcher
• Chouaib Benchraka – doctoral researcher
• Leo Lahti – group leader

---

Learning goals

1. Microbiome research studies interactions between microbes (and human, environment…)
2. Big data requires efficient tools to manipulate the data
3. miaverse is a SummarizedExperiment framework for microbiome analytics

Figure source: Moreno-Indias et al. (2021) Statistical and Machine Learning Techniques in Human Microbiome Studies: Contemporary Challenges and Solutions. Frontiers in Microbiology 12:11.

Motivation

Microbiome research

• Microbiome is a composition of microbes in well-defined area (gut, skin, mouth…)
• Bilateral interaction between human and microbiome –> affects both health and disease.
• The research is based on sequencing (characterization of genes and species).
• Nowadays, multiomics approach is more common (integration of taxonomy information with metabolite data, for example)
• Computational methods are the new microscope
• The research has expanded rapidly in previous years

# Plot publication graph
path <- "data/PubMed_Timeline_Results_by_Year.csv"
df <- read.csv(path, skip = 1)

x <- "Year"
y <- "Count"

plot <- ggplot(df, aes(x = .data[[x]], y = .data[[y]])) +
    geom_bar(stat="identity")
plot

PubMed publications per year with a search term ‘microbiome’ (fetched: Sep 5, 2023)

Big data

• Most of the cohort studies require multiple types of data, for instance microbiome, phenotype and other omics data.
• Linking these types of data could present multiple challenges; to mention:
  • Different sources the data is coming form, could use different types of key identifier of the samples or subjects of the study; which could poses a problem in some cases for linking the data together.
  • The choice of the data structure to combine and store these mulitple types of data is critical as well, in terms of the management, handling and wrangling of the data; to be able to perform the analysis.
• Cohort datasets are large in size, which mostly one cannot conduct an analysis using a normal computer with limited hardware; HPC and cloud computing services are required in this case.
• Last but not least, a quality check reporting of the constructed object is substantially important to verify the correctness of the building process.

Regarding the choice of the data strcuture to handle cohorts with muliple data types or sources MultiAssayExperiment, is considered to be very liable in such situation, easing-up the management and wrangling of the data; once is construced correctly. Moreover, several R packages frameworks are increasingly integrating MultiAssayExperiment and the SummarizedExperimentclass, to provide users with a reliable and ease of use data structure.

miaverse (MIcrobiome Analysis)

• A framework for microbiome analytics
  • mia (analysis)
  • miaViz (visualization)
  • miaSim (simulation)
  • Orchestrating Microbiome Analysis, OMA (tutorial book)
• Based on TreeSummarizedExperiment (TreeSE) class
  • Supports hierarchical data
  • phyloseq class is a subset of TreeSE
  • Extension of SingleCellExperiment class
  • miaverse is compatible with other SummarizedExperiment frameworks
• Purpose to offer tools and tips for microbiome data analysis
• Enables development of versatile analytical workflows in microbiome data science
  • Supports multiomics (MultiAssayExperiment class)
  • Scalable
  • Standardized

The structure of the TreeSummarizedExperiment (TreeSE) class.

The workflow

Importing the dataset

We get the data from MGnify database. It is a EMBL-EBI’s database for metagenomic data. This large microbiome database can be accessed with MGnifyR package which nowadays support TreeSE. The package will be submitted to Bioconductor’s next release.

We chose this dataset…

As loading takes some time, the dataset is already loaded.

For other available datasets and importing methods, check OMA.

# library(MGnifyR)
# mg <- MgnifyClient()
# 
# analyses <- searchAnalysis(mg, "studies", "MGYS00005128")
# analyses <- searchAnalysis(mg, "studies", "MGYS00000596")

# mae <- getResult(mg, analyses)

library(mia)
data("HintikkaXOData")
MAE <- HintikkaXOData
MAE

## A MultiAssayExperiment object of 3 listed
##  experiments with user-defined names and respective classes.
##  Containing an ExperimentList class object of length 3:
##  [1] microbiota: TreeSummarizedExperiment with 12706 rows and 40 columns
##  [2] metabolites: TreeSummarizedExperiment with 38 rows and 40 columns
##  [3] biomarkers: TreeSummarizedExperiment with 39 rows and 40 columns
## Functionality:
##  experiments() - obtain the ExperimentList instance
##  colData() - the primary/phenotype DataFrame
##  sampleMap() - the sample coordination DataFrame
##  `$`, `[`, `[[` - extract colData columns, subset, or experiment
##  *Format() - convert into a long or wide DataFrame
##  assays() - convert ExperimentList to a SimpleList of matrices
##  exportClass() - save data to flat files

As we can see, our data contains three different types of data: microbiota, metabolites and biomarkers; which are all of the TreeSummarizedExperiment class combined in ourMultiAssayExperiment object.

Microbiota:

MAE[["microbiota"]] # or MAE[[1]]

## class: TreeSummarizedExperiment 
## dim: 12706 40 
## metadata(0):
## assays(1): counts
## rownames(12706): GAYR01026362.62.2014 CVJT01000011.50.2173 ... JRJTB:03787:02429 JRJTB:03787:02478
## rowData names(7): Phylum Class ... Species OTU
## colnames(40): C1 C2 ... C39 C40
## colData names(0):
## reducedDimNames(0):
## mainExpName: NULL
## altExpNames(0):
## rowLinks: NULL
## rowTree: NULL
## colLinks: NULL
## colTree: NULL

Metabolites:

MAE[["metabolites"]] # or MAE[[2]]

## class: TreeSummarizedExperiment 
## dim: 38 40 
## metadata(0):
## assays(1): nmr
## rownames(38): Butyrate Acetate ... Malonate 1,3-dihydroxyacetone
## rowData names(0):
## colnames(40): C1 C2 ... C39 C40
## colData names(0):
## reducedDimNames(0):
## mainExpName: NULL
## altExpNames(0):
## rowLinks: NULL
## rowTree: NULL
## colLinks: NULL
## colTree: NULL

Biomarkers:

MAE[["biomarkers"]] # or MAE[[3]]

## class: TreeSummarizedExperiment 
## dim: 39 40 
## metadata(0):
## assays(1): signals
## rownames(39): Triglycerides_liver CLSs_epi ... NPY_serum Glycogen_liver
## rowData names(0):
## colnames(40): C1 C2 ... C39 C40
## colData names(0):
## reducedDimNames(0):
## mainExpName: NULL
## altExpNames(0):
## rowLinks: NULL
## rowTree: NULL
## colLinks: NULL
## colTree: NULL

Additionally, the MAE contains phenotype or metadata about samples or subjects, that could be accessed as follows:

colData(MAE) %>% head %>% knitr::kable()

	Sample	Rat	Site	Diet	Fat	XOS
C1	C1	1	Cecum	High-fat	High	0
C2	C2	2	Cecum	High-fat	High	0
C3	C3	3	Cecum	High-fat	High	0
C4	C4	4	Cecum	High-fat	High	0
C5	C5	5	Cecum	High-fat	High	0
C6	C6	6	Cecum	High-fat	High	0

Exploring the microbiome data:

Taxa Counts

library(dplyr)
rowData(MAE[[1]]) %>% as_tibble() %>% summarise_all(n_distinct) %>% 
    knitr::kable()

Phylum	Class	Order	Family	Genus	Species	OTU
14	23	43	74	225	319	12706

Top 10 Species

library(miaViz)
# we agglomerate the data from OTUs to species,
# and store it at the alternative experiment slot 
altExp(MAE[[1]], "Species") <- agglomerateByRank(MAE[[1]], rank="Species",
                                                 onRankOnly=TRUE, na.rm = FALSE)
# we transform data from counts to relative abundance
altExp(MAE[[1]], "Species") <- transformAssay(altExp(MAE[[1]], "Species"),
                                          assay.type = "counts",
                                          method = "relabundance")
plotAbundanceDensity(altExp(MAE[[1]], "Species"), assay.type = "relabundance",
                     n=10L, point_size=0.5, layout="density") +
    scale_x_log10() +
    labs(title = "Top 10 Species", x="log10(relabundance) ")+
    theme(plot.title = element_text(hjust = 0.5, size = 14),
          text=element_text(size = 12))

Top 10 Genera

# we agglomerate the data from OTUs to species,
# and store it at the alternative experiment slot 
altExp(MAE[[1]], "Genus") <- agglomerateByRank(MAE[[1]], rank="Genus",
                                                 onRankOnly=TRUE, na.rm = FALSE)
# we transform data from counts to relative abundance
altExp(MAE[[1]], "Genus") <- transformAssay(altExp(MAE[[1]], "Genus"),
                                          assay.type = "counts",
                                          method = "relabundance")
plotAbundanceDensity(altExp(MAE[[1]], "Genus"), assay.type = "relabundance",
                     n=10L, point_size=0.5, layout="density") +
    scale_x_log10() +
    labs(title = "Top 10 Genus", x="log10(relabundance) ")+
    theme(plot.title = element_text(hjust = 0.5, size = 14),
          text=element_text(size = 12))

Alpha diveristy

Lets look at ShannonIindex and Observed Richness, where the results are stored back automatically at the colData of the MAE microbiome Species alternative experiment:

altExp(MAE[[1]], "Species") <- estimateDiversity(altExp(MAE[[1]], "Species"),
                                                 index="shannon", 
                                                 assay.type="counts")
altExp(MAE[[1]], "Species") <- estimateRichness(altExp(MAE[[1]], "Species"),
                                                index="observed",
                                                assay.type="counts")
colData(altExp(MAE[[1]], "Species")) %>% head() %>% knitr::kable()

	shannon	observed
C1	1.685654	79
C2	1.368910	55
C3	1.101818	69
C4	1.182855	69
C5	1.189534	61
C6	1.368236	61

Let’s look at the histogram of both indices:

colData(altExp(MAE[[1]], "Species")) %>% as_tibble() %>%
    tidyr::pivot_longer(cols=c(shannon,observed), names_to = "alpha_type",
                        values_to = "estimate") %>% 
    ggplot(aes(x=estimate)) +
    geom_histogram(color = "black", fill = "gray", bins = 30)+
    facet_wrap(~alpha_type, nrow = 1, scales = "free")+
    labs(title = "Alpha Diversity")+
    theme(plot.title = element_text(hjust = 0.5))

Beta diveristy

Let’s look at beta diversity by performing Principal Coordinate Analysis (PCoA) using Bray-Curtis as a distance method on the OTU level data:

library(scater)
MAE[[1]] <- runMDS(MAE[[1]], FUN = vegan::vegdist, name = "PCoA_BC",
              assay.type = "counts")
# Calculating explained variance
e <- attr(reducedDim(MAE[[1]], "PCoA_BC"), "eig");
rel_eig <- e/sum(e[e>0])
# Ploting the ordination 
plotReducedDim(MAE[[1]], "PCoA_BC") + 
    labs(x = paste("PC1 (", round(100 * rel_eig[[1]],1), "%", ")", sep = ""),
         y = paste("PC2 (", round(100 * rel_eig[[2]],1), "%", ")", sep = ""),
         title="PCoA with Bray-Curtis")+
    theme(plot.title = element_text(hjust = 0.5))

With could visualize the same plot using a metadata varible (from colData), to overlay it on our PCoA plot:

# we first assign our colData from MAE to MAE [[1]]
colData(MAE[[1]]) <- colData(MAE)[colnames(MAE[[1]]),]
plotReducedDim(MAE[[1]], "PCoA_BC", colour_by="Fat") + 
    labs(x = paste("PC1 (", round(100 * rel_eig[[1]],1), "%", ")", sep = ""),
         y = paste("PC2 (", round(100 * rel_eig[[2]],1), "%", ")", sep = ""),
         title="PCoA with Bray-Curtis")+
    theme(plot.title = element_text(hjust = 0.5))

Exploring the microbiome and metabolites data:

As an example of analyzing different Omics data stored at the MAE, let’s look at the cross-correlation between microbiome Species and metabolite data:

# First we transform our microbiome Species data with log10
altExp(MAE[[1]], "Species") <- transformAssay(altExp(MAE[[1]], "Species"),
                                              assay.type = "counts",
                                              method = "log10", pseudocount = 1)
# Then we cross-correlae our microbiome data with the metabolites
correlations <- testExperimentCrossCorrelation(MAE,
                                               altexp1="Species",
                                               experiment2 = 2,
                                               assay.type1 = "log10", 
                                               assay.type2 = "nmr",
                                               method = "spearman", 
                                               p_adj_threshold = NULL,
                                               cor_threshold = NULL,
                                               # Remove when mia is fixed
                                               mode = "matrix",
                                               sort = TRUE,
                                               show_warnings = FALSE)

Then with the help of the “ComplexHeatmap” package we visualize the results as a heatmap:

library(ComplexHeatmap) 
plot <- Heatmap(correlations$cor,
                # Print values to cells
                cell_fun = function(j, i, x, y, width, height, fill) {
                    # If the p-value is under threshold
                    if( !is.na(correlations$p_adj[i, j]) & correlations$p_adj[i, j] < 0.05 ){
                        # Print "X"
                        grid.text(sprintf("%s", "X"), x, y, gp = gpar(fontsize = 8, col = "#1dff00"))
                        }
                    },
                heatmap_legend_param = list(title = "", legend_height = unit(5, "cm")),
                column_names_gp = grid::gpar(fontsize = 8),
                row_names_gp = grid::gpar(fontsize = 8),
                column_names_centered=TRUE
                )
plot

Exploring the microbiome and biomarkers data:

Once again we could look at the cross-correlation between microbiome Species and biomarkers data:

# Then we cross-correlae our microbiome data with the biomarkers
correlations <- testExperimentCrossCorrelation(MAE,
                                               altexp1="Species",
                                               experiment2 = 3,
                                               assay.type1 = "log10", 
                                               assay.type2 = "signals",
                                               method = "spearman", 
                                               p_adj_threshold = NULL,
                                               cor_threshold = NULL,
                                               # Remove when mia is fixed
                                               mode = "matrix",
                                               sort = TRUE,
                                               show_warnings = FALSE)

The resulting heatmap:

library(ComplexHeatmap) 
plot <- Heatmap(correlations$cor,
                # Print values to cells
                cell_fun = function(j, i, x, y, width, height, fill) {
                    # If the p-value is under threshold
                    if( !is.na(correlations$p_adj[i, j]) & correlations$p_adj[i, j] < 0.05 ){
                        # Print "X"
                        grid.text(sprintf("%s", "X"), x, y, gp = gpar(fontsize = 8, col = "#1dff00"))
                        }
                    },
                heatmap_legend_param = list(title = "", legend_height = unit(5, "cm")),
                column_names_gp = grid::gpar(fontsize = 8),
                row_names_gp = grid::gpar(fontsize = 8),
                column_names_centered=TRUE
                )
plot

Thanks!

• more material in OMA
• miaverse logo
• project website and QR also.
• Contact info
• Poster info

Session info

sessionInfo()

## R version 4.3.0 (2023-04-21)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 20.04.6 LTS
## 
## Matrix products: default
## BLAS/LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.8.so;  LAPACK version 3.9.0
## 
## locale:
##  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=fi_FI.UTF-8        LC_COLLATE=en_US.UTF-8     LC_MONETARY=fi_FI.UTF-8    LC_MESSAGES=en_US.UTF-8   
##  [7] LC_PAPER=fi_FI.UTF-8       LC_NAME=C                  LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=fi_FI.UTF-8 LC_IDENTIFICATION=C       
## 
## time zone: Europe/Helsinki
## tzcode source: system (glibc)
## 
## attached base packages:
## [1] grid      stats4    stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
##  [1] knitr_1.43                     ComplexHeatmap_2.17.0          scater_1.29.3                  scuttle_1.11.2                 tidyr_1.3.0                   
##  [6] dplyr_1.1.2                    miaViz_1.9.0                   ggraph_2.1.0                   ggplot2_3.4.2                  BiocManager_1.30.21.1         
## [11] mia_1.9.9                      MultiAssayExperiment_1.27.0    TreeSummarizedExperiment_2.9.0 Biostrings_2.69.2              XVector_0.41.1                
## [16] SingleCellExperiment_1.23.0    SummarizedExperiment_1.31.1    Biobase_2.61.0                 GenomicRanges_1.53.1           GenomeInfoDb_1.37.2           
## [21] IRanges_2.35.2                 S4Vectors_0.39.1               BiocGenerics_0.47.0            MatrixGenerics_1.13.1          matrixStats_1.0.0             
## 
## loaded via a namespace (and not attached):
##   [1] splines_4.3.0               bitops_1.0-7                ggplotify_0.1.1             tibble_3.2.1                polyclip_1.10-4             DirichletMultinomial_1.43.0
##   [7] lifecycle_1.0.3             rstatix_0.7.2               doParallel_1.0.17           lattice_0.21-8              MASS_7.3-60                 backports_1.4.1            
##  [13] magrittr_2.0.3              sass_0.4.7                  rmarkdown_2.23              jquerylib_0.1.4             yaml_2.3.7                  cowplot_1.1.1              
##  [19] DBI_1.1.3                   RColorBrewer_1.1-3          abind_1.4-5                 zlibbioc_1.47.0             purrr_1.0.1                 RCurl_1.98-1.12            
##  [25] yulab.utils_0.0.6           tweenr_2.0.2                circlize_0.4.15             GenomeInfoDbData_1.2.10     ggrepel_0.9.3               irlba_2.3.5.1              
##  [31] tidytree_0.4.4              vegan_2.6-4                 permute_0.9-7               DelayedMatrixStats_1.23.0   codetools_0.2-19            DelayedArray_0.27.10       
##  [37] DT_0.28                     ggforce_0.4.1               tidyselect_1.2.0            shape_1.4.6                 aplot_0.1.10                farver_2.1.1               
##  [43] ScaledMatrix_1.9.1          viridis_0.6.4               jsonlite_1.8.7              GetoptLong_1.0.5            BiocNeighbors_1.19.0        decontam_1.21.0            
##  [49] tidygraph_1.2.3             iterators_1.0.14            foreach_1.5.2               tools_4.3.0                 ggnewscale_0.4.9            treeio_1.25.2              
##  [55] Rcpp_1.0.11                 glue_1.6.2                  gridExtra_2.3               SparseArray_1.1.11          BiocBaseUtils_1.3.2         xfun_0.39                  
##  [61] mgcv_1.9-0                  withr_2.5.0                 fastmap_1.1.1               bluster_1.11.4              fansi_1.0.4                 digest_0.6.33              
##  [67] rsvd_1.0.5                  R6_2.5.1                    gridGraphics_0.5-1          colorspace_2.1-0            Cairo_1.6-0                 RSQLite_2.3.1              
##  [73] utf8_1.2.3                  generics_0.1.3              DECIPHER_2.29.0             graphlayouts_1.0.0          htmlwidgets_1.6.2           S4Arrays_1.1.5             
##  [79] pkgconfig_2.0.3             gtable_0.3.3                blob_1.2.4                  htmltools_0.5.5             carData_3.0-5               clue_0.3-64                
##  [85] scales_1.2.1                png_0.1-8                   ggfun_0.1.1                 rstudioapi_0.15.0           reshape2_1.4.4              rjson_0.2.21               
##  [91] nlme_3.1-162                cachem_1.0.8                GlobalOptions_0.1.2         stringr_1.5.0               parallel_4.3.0              vipor_0.4.5                
##  [97] pillar_1.9.0                vctrs_0.6.3                 ggpubr_0.6.0                car_3.1-2                   BiocSingular_1.17.1         beachmat_2.17.15           
## [103] cluster_2.1.4               beeswarm_0.4.0              evaluate_0.21               magick_2.7.5                cli_3.6.1                   compiler_4.3.0             
## [109] rlang_1.1.1                 crayon_1.5.2                ggsignif_0.6.4              labeling_0.4.2              plyr_1.8.8                  ggbeeswarm_0.7.2           
## [115] stringi_1.7.12              viridisLite_0.4.2           BiocParallel_1.35.3         munsell_0.5.0               lazyeval_0.2.2              Matrix_1.6-0               
## [121] patchwork_1.1.2             sparseMatrixStats_1.13.0    bit64_4.0.5                 highr_0.10                  igraph_1.5.0.1              broom_1.0.5                
## [127] memoise_2.0.1               bslib_0.5.0                 ggtree_3.9.0                bit_4.0.5                   ape_5.7-1